[lsst-dm-stack-users] Photographic archive in Stripe 82

[Dave Monet]

Thanks to Mario for setting up the DM-STACK mailing list.

I have yet to try downloading and making the stack, but this will be
a high priority once I get back from my next trip.

I just wanted to pass along a possibly relevant Action that I took
from the AHM.  As folks may or may not know, I wasted 20 years of
my life digitizing the photographic archive from the great Schmidt
and Astrograph surveys.  Given the current DM interest in Stripe 82,
I have gone back and fed the 287 plates of that region that we scanned
to SExtractor, and will be trying study the proper motions in the region.
Schmidt plates are 6.5 degrees across, so the 287 plates in the region
of 300 <= RA <= 60 suggests something like 10 visits per piece of sky.

I will be amused (and many other things) to see the difference in
astrometric quality between the SExtractor centroids and the ones
that I computed.  I am pretty sure that I do not have the correct
version of the SExtractor CONFIG and PARAM files, but the re-processing
takes less than a week so I can redo it as necessary.

For those too young or who never learned, our Precision Measuring Machine
(PMM) used a pair of 1394x1037 cameras to image the plates.  These generated
8-bit images that were processed in real time to make our catalogs.  The
pixel scale is close to 0.92 arcsec/pixel, but rest assured that the
images have reasonable sampling because of the photographic process.
588 footprints were needed to scan each plate.  The metrology of
each footprint is defined by the plate+optics+CCD and not by the motion
of the measuring stage.  Hence, there is reason to hope that the 
differential
astrometric accuracy for brighter stars in a single footprint could be
in the range of 0.1 arcsec or even better.  It was the stitching of the
footprints together and the solution for (RA,Dec) that degraded the
accuracy of the positions that appeared in the USNO-A and -B catalogs.

I have attached (I think) the ASCII file of the plate metadata.  It can
be loaded into MySQL using something like
      DROP TABLE IF EXISTS plates82;
      CREATE TABLE plates82 (
              observatory CHAR(3),
              prefix CHAR(2),
              plate INT,
              mjd DOUBLE,
              r0 DOUBLE,
              d0 DOUBLE,
              wave DOUBLE,
              filter CHAR(1)
              );
      LOAD DATA LOCAL INFILE 'stripe82.asc' INTO TABLE plates82;

I will be processing each footprint with respect to my reference SDSS
catalog (from DR7) as well as the DM version once I learn how to download
and use it.  In some small way, this will assist the debugging of the
differential astrometry pipeline that might be used for LSST data.  It
will be amusing to see if I can tell the astrometric difference between
Stripe 82 reductions (DRx, DM, etc.).

I can make images, catalogs, or solutions available to those with more
spare time than brain cells.  The images and catalogs should be done
today, and the astrometric solutions by the end of the month.

-Dave

PS
For the curious, the PMM digitized the entire POSS-I and POSS-I archives,
and not just the plates accepted for the Survey.  We also digitized the
Luyten and PIMWA plates.  It is also the case that the Northern and
Southern surveys included the band at Dec=0, and we digitized all plates
from all surveys except the ESO-B.  PMM has died of old age and cannot
be revived.  We must live with the decisions made over 15 years ago.
    observatory:
        PAL is the Palomar Oschen Schmidt
        AAO is for the UK Schmidt
    prefix:
        PO is POSS-I "blue"
        PE is POSS-I "red"
        PJ is POSS-II "blue"
        PF is POSS-II "red"
        PN is POSS-II "IR"
        PL is used for the Luyten plates of any color
        AB is the SERC-J "blue"
        AR is the SERC-R + -ER "red"
        AN is the SERC-I "IR"
    mjd:
        The nominal MJD for the middle of the exposure.
    r0,d0
        The nominal boresight in J2000 coordinates in degrees.  There is
        error here because some surveys used the B1950 coordinates of date
        and others used the epochless nominal pointing.  One should refer to
        the individual footprint files for the useful J2000 coordinates.
    wave:
        The nominal mid-point of the passband in microns.
    filter:
        The O, E, J, F, N of the nominal emulsion+filter combination.
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: stripe82.asc
URL: <https://listserv.lsstcorp.org/pipermail/dm-users/attachments/20120907/e9b95225/attachment.asc>